# Websites make promises to machines. We help keep them true.

Search engines, AI assistants, ad buyers and security researchers read a handful of small public files before they read a single page. Flowpane exists so those files say what their owners mean, today and every day after.

## Even the best-run websites lose track.

Five public files from household names, checked on 3 October 2026. None of these sites is broken. That is the point: when one of these files drifts, nothing tells anyone.

- **Stale: A national broadcaster** (`sitemap.xml`). Its main sitemap listed six others. Three were made for the 2014 elections, and one of those held nothing but template code that was never processed. `<loc>…/events/vote2014/sitemap.xml</loc>`
- **Out of date: A global payments company** (`robots.txt`). Listed 88 bots to block, most from the early 2000s and one identified as Windows 95, yet named no AI crawler. Opened in a browser, it downloaded instead. `User-agent: Mozilla/4.0 (compatible; BullsEye; Windows 95)`
- **Silent: A second payments network** (`robots.txt`). Served an empty robots.txt on its main domain. Silence is a policy too: every crawler, AI included, may go everywhere. `GET /robots.txt → 200 OK, 0 bytes`
- **Inconsistent: A resale marketplace** (`robots.txt`). Stated its AI policy in two competing vocabularies, added three directives no crawler reads, and its comments forbade what one of its rules allowed. `Content-Signal: ai-train=no, search=yes, ai-input=yes`
- **Good: A travel platform** (`robots.txt`). Kept AI training crawlers out, let AI search and assistants in, and greeted readers with an ASCII logo. Even here, one AI name it listed had been retired. `# Hello bot, engineer, or very lost layperson!`

Names left out on purpose. We keep a dated copy of every file quoted here.

## Our mission

Every website understood the way its owners intend, by everyone and everything that reads it.

We measure that in outcomes you can see, not in how many warnings we can produce.

- **Found, as intended.** The pages you want in search are crawlable and listed, and the ones you want kept out stay out.
- **One AI policy.** robots.txt, ai.txt and llms.txt say the same thing about AI, so every crawler gets one answer.
- **Reachable when it matters.** A researcher who finds a security problem reaches a person, through a contact that has not expired.
- **Revenue protected.** Ad buyers can verify who may sell your inventory, so bids are not lost to a stale ads.txt.
- **Proof on request.** When a client, auditor or regulator asks what your site said and when, you have the record.
- **Time back.** Your team stops checking files by hand, site after site, and hears about changes without going looking.

## A watchdog, not a replacement.

Your CMS, your SEO plugin and your CDN write these files, and tools like WordPress, Yoast, Rank Math and Cloudflare do it well. Flowpane does not replace any of them, and it does not change your CDN. It watches what they publish, checks it against everything else your site says, and tells you when something changes.

- Written by: Your CMS, SEO plugins, Your CDN, Developers and deploys.
- Published as: robots.txt, sitemap.xml, security.txt, ads.txt, llms.txt.
- Read by: Search engines, AI crawlers and assistants, Ad buyers, Security researchers.

Flowpane watches what is published, whoever wrote it.

- **Not an SEO agency.** We do not promise rankings or sell traffic. Search engines are one of the readers we check for, and your SEO work stands on firmer ground.
- **Not another generator.** Plenty of tools write these files, and write them well. Our job is what comes after: checking that what was published is true, agrees with everything else, and stays that way.
- **Reads first.** Flowpane checks from the outside and changes nothing on its own. With a connected CMS, it applies only the fixes an admin has approved, where the platform allows.

## Principles you can check.

- **We never sign in to your site.** Checks run from the outside, over public HTTPS, from a crawler that says who it is. [How Flowpane fetches](https://flowpane.com/learn/how-flowpane-fetches.md)
- **We practise what we check.** Our own robots.txt, sitemap, security.txt and llms.txt are public. Read them, and hold us to them. [Read our robots.txt](https://flowpane.com/robots.txt)
- **Product data stays product data.** This website and the Flowpane app keep separate databases, and app data is never used for advertising. [Privacy notice](https://flowpane.com/privacy.md)
- **Plain language, legal pages included.** Our policies are written to be read, and they sit together in one place. [Legals](https://flowpane.com/legal.md)
- **Pricing you can read in one sitting.** Every plan checks the same files with the same rules. Plans differ in scale, cadence and how long the record is kept. [Pricing](https://flowpane.com/pricing.md)
- **A person reads every message.** Every enquiry lands in one place and is routed to the right person. [Contact](https://flowpane.com/contact.md)

## Independent, and focused on one job.

Flowpane builds one product and gives it our full attention. We publish when there is something worth saying, read every message, and keep our own files in order.

- [Contact](https://flowpane.com/contact.md): Enterprise, partnerships and press.
- [Careers](https://flowpane.com/careers.md): Open roles, when there are any.
- [Press](https://flowpane.com/press.md): Boilerplate, product facts and brand assets.

The Flowpane team